基于石墨烯的触觉传感器及三维力解耦研究

发布时间：2018-11-10 22:39

【摘要】：智能化是机器人技术最新发展的一个重要方向,而实现智能化的一个关键技术——触觉传感器技术,是制约其发展的重要因素。对于人类触觉的模拟,既要满足三维力测量的高灵敏度,又要有人类皮肤的柔性。目前针对强力信号(10 kPa)的检测技术已经成熟,而对于弱力(5 kPa)及三维力信号的检测仍需研究,同时柔性传感器器件制造水平仍需提高。鉴于此,本课题拟采用阵列结构化的石墨烯薄膜作为力学敏感材料,通过优化设计传感器的结构和制造工艺,制备出高性能力学传感器,并采用一体化的信号采集系统,实现力学信号的快速采集与分析。利用材料的界面压阻效应,设计了一种基于石墨烯的柔性三维力触觉传感器。传感器总体结构包括表面凸起层、压阻层和柔性电极三个部分。对界面电阻产生的机理进行了分析,研究了传感器三维力检测原理。进一步利用硅光刻技术、湿法刻蚀工艺、反拷铸膜法和层层自组装技术(Layer-By-Layer,LBL)制作了带有石墨烯薄膜的压阻层。设计并制备出了柔性电极层,将各部分组装成柔性三维力传感器的压阻单元;分别对其进行了正压力与任意三维力的检测性能测试。当接触压力小于500 Pa时,传感器灵敏度为-1.71KPa-1,500~5000 Pa时,传感器灵敏度为-0.0178 KPa-1;响应和回复时间分别为7.5ms和7.4 ms;传感器可实现对正压力、剪切力、空间任意方向力的快速和高灵敏度检测,其最低检测限为2 Pa,且具有良好的稳定性。进一步搭建了阵列触觉传感器的信号采集电路,首先利用分压的方式设计了信号检出电路,然后通过高速选通芯片CD4051对每个阵列传感单元的输出信号进行采集。利用STM32单片机的A/D模数转换器实现信号由模拟量到数字量转换,使用STM32单片机对数字信号处理分析,最后将输出信号通过PC显示出来。最后对触觉传感器进行了耦合分析,分析结果表明传感器的耦合计算结果受到设计原理、测量方式和传感器制造装配精度的影响,因此,每个方向的输出信号会受到多个外界因素的干扰。鉴于此,采用了BP神经网络对触觉传感器进行了解耦研究。解耦计算结果的最大误差为4.3%FS,最后编写了解耦界面。
[Abstract]:Intellectualization is an important direction in the latest development of robot technology, and tactile sensor technology, which is a key technology to realize intelligentization, is an important factor restricting the development of robot technology. For human tactile simulation, it is necessary to satisfy the high sensitivity of 3D force measurement and the flexibility of human skin. At present, the detection technology of strong signal (10 kPa) is mature, but the detection of weak force (5 kPa) and 3D force signal still needs to be studied, and the manufacturing level of flexible sensor device still needs to be improved. In view of this, this subject intends to use the graphene thin film of array structure as the mechanical sensitive material, by optimizing the structure and manufacturing process of the sensor, the high performance mechanical sensor is prepared, and the integrated signal acquisition system is adopted. The quick acquisition and analysis of mechanical signals are realized. A flexible 3D force tactile sensor based on graphene was designed based on the interfacial piezoresistive effect. The overall structure of the sensor consists of three parts: surface protruding layer, piezoresistive layer and flexible electrode. The mechanism of interface resistance is analyzed, and the three-dimensional force detection principle of sensor is studied. The piezoresistive layer with graphene film was prepared by silicon lithography, wet etching, reverse coating casting and layer self-assembly (Layer-By-Layer,LBL). The flexible electrode layer is designed and fabricated, and the piezoresistive elements of the flexible three-dimensional force sensor are assembled, and the measurement performance of the positive pressure and arbitrary three-dimensional force are tested respectively. When the contact pressure is less than 500 Pa and the sensitivity of the sensor is-1.71KPa-1500~5000 Pa, the sensitivity of the sensor is-0.0178 KPa-1; response and recovery time is 7.5ms and 7.4 ms;, respectively. The sensor can detect the positive pressure, shear force and spatial force in any direction quickly and with high sensitivity. The minimum detection limit is 2 Pa, and the sensor has good stability. Furthermore, the signal acquisition circuit of the array tactile sensor is built. Firstly, the signal detection circuit is designed by using the method of partial voltage, and then the output signal of each array sensor unit is collected by high speed gated chip CD4051. The A / D converter of STM32 microcontroller is used to realize the signal conversion from analog to digital, and the digital signal is processed and analyzed by STM32 single chip computer. Finally, the output signal is displayed through PC. Finally, the coupling analysis of the tactile sensor is carried out. The results show that the coupling calculation results of the sensor are affected by the design principle, the measurement method and the assembly accuracy of the sensor. The output signal in each direction is disturbed by multiple external factors. In view of this, BP neural network is used to decouple the tactile sensor. The maximum error of the decoupling results is 4.3 FSs. Finally, the decoupling interface is written.
【学位授予单位】：哈尔滨工业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP212

【参考文献】